{"id":"W1978121157","doi":"10.1016/j.orggeochem.2004.01.009","title":"Definition and characterization of petroleum compositional families in Williston Basin, North America using principal component analysis","year":2004,"lang":"en","type":"article","venue":"Organic Geochemistry","topic":"Hydrocarbon exploration and reservoir analysis","field":"Engineering","cited_by":35,"is_retracted":false,"has_abstract":false,"ca_institutions":"Geological Survey of Canada","funders":"University of Crete","keywords":"Principal component analysis; Sterane; Compositional data; Biomarker; Geology; Characterization (materials science); Petroleum; Gasoline; Multivariate statistics; Boiling point; Structural basin; Environmental science; Hopanoids; Source rock; Mineralogy; Chemistry; Statistics; Mathematics; Paleontology; Materials science; Organic chemistry","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00023231,0.0001514785,0.0001809886,0.003621166,0.0007875372,0.0007939455,0.0003211501,0.0001048348,0.0006326103],"category_scores_gemma":[0.00099717,0.0001534002,0.0001462493,0.002748447,0.0004832294,0.0003841348,0.00077801,0.0001555902,0.0001207267],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001541251,"about_ca_system_score_gemma":0.00239006,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.4965326,"about_ca_topic_score_gemma":0.7153265,"domain_scores_codex":[0.99984,0.00002615451,0.000016709,0.00004138449,0.00004683291,0.00002884832],"domain_scores_gemma":[0.9993446,0.00009815306,0.0001028494,0.00003751582,0.0003520894,0.00006483204],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00008579824,0.00003389297,0.9565447,0.00003154165,0.00005810329,0.000100091,0.001717974,0.000686422,0.005423307,0.0007414686,0.0004712451,0.03410538],"study_design_scores_gemma":[0.000002184623,0.00000736807,0.9944916,0.000009076937,0.00001098429,0.00004219705,0.001522136,0.001775511,0.0005227182,0.0001688499,0.001442965,0.000004445111],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9965867,0.0001243131,0.0008525983,0.00002416328,0.000001327827,0.0000215603,0.0005517592,0.00001258039,0.001825081],"genre_scores_gemma":[0.9953994,0.0001372351,0.002640952,0.00001325543,0.00000167799,0.00003020859,0.0009427923,0.000009078225,0.0008254172],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4965326,"threshold_uncertainty_score":0.9872848,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.007972012608981748,"score_gpt":0.1896957195235431,"score_spread":0.1817237069145613,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}